Intelligent model and optimization of ultrasound-assisted extraction of antioxidants and amylase enzyme from Gnaphalium affine D. Don

Abstract

The study uses ultrasound-assisted extraction to recovery the antioxidant and amylase enzyme from Gnaphalium affine D. Don, namely “chewcut” in Thailand. The study involves two statistical methods: artificial neural networks (ANN) and response surface methodology (RSM) to model and optimize extraction procedure for improving the yield of antioxidant and amylase enzyme activity (AEA). Both RSM and ANN showed the potential to predict and find the optimal extraction conditions. However, ANN model could give more accurate values compared with validation test. ANN model found that under optimal conditions (temperature: 65.92 °C, ultrasonic power: 58.22 %, extraction time: 37.95 min), the total phenolic compounds, total flavonoid compounds, antioxidant activity and AEA were 218.35 ± 0.34 mgGAE/g, 0.554 ± 0.045 mgQE/g, 84.2 ± 0.2 %, 364.14 ± 1.35 mg-maltose/g. This is the first report on amylase potential of chewcut, which could be further served as the natural enzyme source. Moreover, by adding its bioactive compounds, it may be possible to improve nutraceutical properties and quality of products.